PROJECT SUMMARY ? Project 1 The 2016 US Preventive Services Task Force and 2015 American Cancer Society guidelines call for reduced screening intensity and for providers to discuss with women their preferences for screening. This push for individualized decision-making has triggered substantial movement toward risk-based screening utilization that does not rely solely on a woman's age to determine when to start and stop screening or screening frequency. Yet, the explicit incorporation of risk assessment into screening guidelines for the general population is in its early stages. Risk prediction is a key aspect of risk-based screening; available models predict breast cancer risk overall, but do not take into account the heterogeneity of breast cancer biology or the ability of screening modalities to detect breast cancer. Project 1 aims to advance a new risk-based screening paradigm that identifies women's absolute cumulative risk of screening detection, failures, and false alarms while explicitly considering screening interval and modality used. Aim 1) Using the existing and expanded Breast Cancer Surveillance Consortium (BCSC) infrastructure, project investigators will develop new risk prediction models to identify women at high risk of: 1) early-stage screen-detected cancer (stage I/IIa invasive cancer), 2) screening failure (interval invasive cancer or screen-detected stage IIb or higher), and 3) false alarms (false-positive tests and benign biopsies). Using BCSC data on over 1,000,000 women aged 40-79 years undergoing digital mammography or tomosynthesis, and over 13,000 invasive breast cancers identified, the project team will identify clinical and imaging factors associated with breast cancer risk that impact screening outcomes, generating six-year risk estimates for each screening outcome for annual, biennial, and triennial screening regimens. Aim 2) The project team will identify actionable levels of clinically-meaningful risk of screening failure and false alarm, and associated long-term screening outcomes using women and clinician surveys and Delphi panels. Aim 3) Three established breast cancer simulation models will be used to evaluate risk-targeted digital and tomosynthesis screening in which starting/stopping ages and intervals are based on actionable risk levels from Aim 2, in comparison to current age-based guidelines. Model outputs will assess the long-term benefits (breast cancer deaths averted and life years gained), false alarms, and costs of each strategy. This proposal will answer: 1) Which women are at high risk of screening failures and require an alternative strategy to biennial mammography, and which women are at low risk and can be screened less often? 2) What risk levels of poor screening outcomes impact women and clinician preferences for screening strategies? 3) Does a risk-based screening strategy with tomosynthesis improve the effectiveness of screening compared to an age- based strategy? This project will provide evidence to guide women, health care providers, and policymakers on screening strategies based on short-term risk of screening failures and false alarms, and associated long-term outcomes, with the goal to improve early detection of aggressive tumors while minimizing harms.